Yes. Anthropic's announcement says it plans to expand its use of Google TPUs, "including up to one million TPUs", bringing "well over a gigawatt of capacity online in 2026."
A multi-chip strategy
The same announcement describes a multi-platform approach that "efficiently uses three chip platforms: Google's TPUs, Amazon's Trainium, and NVIDIA's GPUs."
So one of the best-known AI labs spreads work across three accelerator families rather than picking a single winner. Each has a different software stack: TPUs run through XLA and TorchTPU, and Trainium runs through the Neuron SDK.
For the chips themselves, see what Ironwood is and NVIDIA's main competitors in AI.
